Summary: Meta Overhauls AI Infrastructure to Keep Pace with Innovation

Meta, formerly known as Facebook, has been slow to adopt expensive AI-friendly hardware and software systems, which has led to a significant gap in its ability to develop AI. Despite high-profile AI research investments, the platform required an overhaul to boost its AI infrastructure. This has led to the company spending roughly $4 billion more per quarter on capital expenditures, according to company disclosures. The overhaul included leadership changes, a scrapped AI chip project and Meta’s shift towards GPUs. The company is now developing its own in-house chip capable of training AI models and performing inferences. The restructuring has also impacted Meta’s capacity to develop generative AI products, which are driving an arms race among tech giants. Carnegie Mellon University Professor Kris Kitani said companies that don’t invest in AI infrastructure now will struggle to catch up later.

Related articles

Inside Meta’s scramble to catch up on AI

As the summer of 2022 came to a close, Meta CEO Mark Zuckerberg gathered his top lieutenants for a dissection of the company’s computing capacity.

Read the complete article at: www.reuters.com